Acoustic wave separation (AWS) is the next-generation technique for continuous clarification of Chinese Hamster Ovary (CHO) cells. In this study, a hybrid strategy coupling a mechanistic and a data-driven model was used to investigate the process control dynamics and the clarification efficiency of CHO cells using Cadence<sup>TM</sup> Acoustic Separator (CAS). The CAS constitutes of four acoustic chambers connected in series, where each chamber is preceded by a dedicated online turbidity meter. The objective of the mechanistic model is to predict the cell separation efficiency based on the incoming load cell density, while the empirical model predicts the protein recovery in the clarified supernatant, and both combined as a hybrid model provide a priori estimate of optimum operating conditions for enhanced clarification. The framework works as a Process Analytical Technology (PAT) tool and is linked with the five online turbidity meters, which provide real-time turbidity data of each acoustic chamber inlet. A control algorithm was developed to monitor the turbidity data at regular time intervals, and any deviation in the incoming load cell density at regular intervals was detected in advance. Incorporating a feed-forward control strategy, the error with respect to the model-predictive set point was computed by the algorithm and it implemented proportionate control action in each acoustic chamber before a significant deviation occurred. Experimental validation confirmed that the hybrid model predictive control was able to regulate final turbidity levels and that deviation &lt; 10% was observed compared to the setpoint in multiple cases. Overall, a stricter control was achieved, resulting in both protein recovery and cell separation efficiency &gt; 85%. The study alludes to the merit of hybrid models for model-based control and depicts a potential application in integrating AWS in continuous processing.

Hybrid model predictive control strategy of CHO cell acoustic clarification and its integration in continuous manufacturing

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.